Abstract
Traditional angle sensors have problems such as insufficient accuracy and complex installation and debugging when measuring the deflection angle of tractor guide wheels，which hinder the high precision and stability requirements of tractor automatic navigation operations. Therefore，this paper aims to construct and validate a novel system based on the principles of Global Navigation Satellite System（GNSS）and Inertial Navigation System（INS），which is capable of real-time calculation of dynamic angles of the tractor's steering wheel. First，a in-depth analysis of GNSS/INS fusion principles was conducted，and corresponding GNSS/INS data measurement bench tests were designed to obtain measurement values from angle sensors. Subsequently，three filtering algorithms，namely Kalman Filter（KF），Extended Kalman Filter（EKF），and Unscented Kalman Filter（UKF），were employed to filter the angle measurement values，evaluating metrics such as overshoot，response time，and stability，and the results showed that UKF performed best in this system. Second，steering and straight-line performance tests were conducted under different conditions（asphalt road and farmland）for tractor automatic navigation. In the steering performance tests，fixed measurements at different angles of 1°，5°，and 15° were conducted，and the results showed that the response time of the two environments was the smallest at 1°，but both produced the largest overshoot. At the same time，an increase in angle demonstrated a significant reduction in steady-state error. In the straight-line automatic navigation performance tests，the maximum lateral deviations for asphalt road and farmland were 17.51 mm and 18.52 mm，respectively，with system errors of 10.45 mm（2σ）and 21.07 mm（2σ）. A comparison between the two conditions indicated superior performance metrics on asphalt road，yet both conditions met the requirements of precision，response time，and stability for automatic navigation systems，thus suitable for tractor automatic navigation operations.
